I am on Sling 11, which uses Jackrabbit Oak as content repository. I was wondering how to set up Sling to store the JCR repo on an RDBMS (DB2 to be specific).
I found this link on Jackrabbit Persistence, but looks like it does not apply to Oak and Oak documentation is mostly about MongoDB.
Also found an implementation of a Cassandra Resource Provider, although that seems designed to access specific paths mapped to Cassandra without using Oak.
Thanks,
Answering here but credit goes to Sling user's mailing list
In boot.txt add a new running mode (in my case oak_db2)
[settings] sling.run.mode.install.options=oak_tar,oak_mongo,oak_db2
Download Sling's datasource project and compile it.
In oak.txt configure the running mode (this will load the bundles for you in Felix):
[artifacts startLevel=15 runModes=oak_db2]
com.h2database/h2-mvstore/1.4.196
com.ibm.db2/jcc4/11.1
org.apache.sling/org.apache.sling.datasource/1.0.3-SNAPSHOT
And set-up the services that will manage persistence:
[configurations runModes=oak_db2]
org.apache.jackrabbit.oak.plugins.document.DocumentNodeStoreService
documentStoreType="RDB"
org.apache.sling.datasource.DataSourceFactory
url="jdbc:db2://10.1.2.3:50000/sling"
driverClassName="com.ibm.db2.jcc.DB2Driver"
username="****"
password="****"
datasource.name="oak"